A spatially fractionated radiotherapy dose distortion quantification system based on low-pass filtering modeling

Abstract

The present application relates to the field of dose data frequency domain analysis, and discloses a spatial segmentation radiotherapy dose distortion quantification system based on low-pass filter modeling, the present application proves the mathematical equivalence of the deformation vector field probability weighted cumulative operator by the low-pass filter equivalence analysis model, and performs frequency domain analysis, so that the deviation between the planned dose and the cumulative dose can be attributed and quantitatively explained at the frequency domain level; the peak-valley dose index acquisition module is used to realize accurate quantification of the peak-valley structure distortion mode of lattice radiotherapy; the frequency domain physical prediction model is used to determine the upper limit of the gradual degradation rate and the steepness parameter of the degradation curve according to the characteristic frequency and the cutoff frequency of the dose distribution of the lattice radiotherapy, so as to obtain the predicted peak-valley dose degradation rate, and the dose degradation degree is calculated in advance based on the motion amplitude; the quality determination module is used to automatically output the quality determination result according to the predicted peak-valley dose degradation rate, so as to realize the automation and standardization of dose distortion evaluation.

Description

Technical Field

[0001] This invention relates to the field of frequency domain analysis technology for dose data, and in particular to a spatial segmentation radiotherapy dose distortion quantization system based on low-pass filter modeling. Background Technology

[0002] Spatially Fractionated Radiation Therapy (SFRT), by creating a non-uniform dose distribution of high-dose peaks and low-dose valleys within the tumor, has attracted widespread attention in the radiotherapy of large-volume, locally advanced tumors. Lattice Radiotherapy (LRT), as an important form of spatially fractionated radiotherapy, utilizes techniques such as volumetric intensity-modulated arc therapy to generate regularly arranged spherical high-dose lattice points within the tumor. The core dosimetric characteristic of LRT is the peak-to-valley dose ratio (PVDR), which directly determines the degree to which a non-uniform dose distribution within the target area is achieved.

[0003] Existing lattice radiotherapy dose accumulation assessment and distortion analysis systems are mainly divided into:

[0004] 1. The general dose accumulation calculation system adopts a probability weighted accumulation method based on deformation vector field (DVF). After mapping the dose distribution under multiple respiratory phases to the reference coordinate system, it performs voxel-by-voxel weighted summation and finally outputs the numerical result of the cumulative dose distribution. It only completes the numerical calculation and does not analyze the structural changes of the dose distribution.

[0005] 2. The post-dose verification system assesses the degree of dose distortion by comparing the distribution differences between the planned dose and the cumulative dose after all dose accumulation calculations are completed. The results can only be obtained after the calculations are completed, and it is impossible to assess the risk of dose degradation under different exercise conditions in advance during the planning stage.

[0006] 3. Experience threshold judgment system, which makes a rough judgment based on whether the amplitude of movement exceeds a preset threshold.

[0007] Therefore, the existing technology has the following technical problems:

[0008] 1. The probability weighted cumulative calculation based on the deformation vector field can only output the final numerical result of the cumulative dose. It cannot explain the structural impact of the probability weighted cumulative calculation process of the deformation vector field on the dose distribution from the signal processing level, which leads to the deviation between the planned dose and the cumulative dose being unable to be attributed and quantitatively explained.

[0009] 2. The degree of dose distortion can only be obtained after dose verification, and it is impossible to prospectively quantify it based on the individual patient's range of motion during the planning stage, which limits the ability to predict quality during the planning stage.

[0010] 3. The dose degradation risk assessment relies solely on simple amplitude threshold judgments, lacking precise physical basis and failing to meet the differentiated assessment needs under different amplitude conditions. Summary of the Invention

[0011] The purpose of this invention is to address the technical problems of existing technologies, such as the lack of frequency domain analysis in deformation vector field accumulation calculation, the lack of forward-looking prediction of dose distortion, and the lack of physical model basis for risk assessment, by providing a spatial segmentation radiotherapy dose distortion quantification system based on low-pass filter modeling.

[0012] To achieve the above-mentioned objectives, the embodiments of the present invention provide the following technical solutions:

[0013] A spatial segmentation radiotherapy dose distortion quantification system based on low-pass filter modeling includes a data preprocessing module, a low-pass filter equivalence analytical model, a peak-valley dose index acquisition module, a frequency domain physical prediction model, and a quality judgment module.

[0014] The data preprocessing module is used to read physical dose data and geometric space data, preprocess the physical dose data, and obtain preprocessed physical dose data.

[0015] The low-pass filter equivalence analytical model is used to perform mathematical equivalence proof and frequency domain analysis using preprocessed physical dose data to obtain the cutoff frequency and the characteristic frequency of the lattice radiotherapy dose distribution.

[0016] The peak-valley dose index acquisition module obtains a comprehensive index of peak-valley dose degradation rate based on geometric spatial data and preprocessed physical dose data.

[0017] The frequency domain physical prediction model is used to fit and determine the upper limit of the progressive degradation rate and the steepness parameter of the degradation curve based on the characteristic frequency and cutoff frequency of the lattice radiotherapy dose distribution, and to obtain the predicted peak and valley dose degradation rate.

[0018] The quality assessment module is used to obtain quality assessment results based on the predicted peak and valley dose degradation rates.

[0019] Compared with existing technologies, the beneficial effects of this invention are as follows: By performing mathematical equivalence proof and frequency domain analysis on the probability weighted accumulation operator of the deformation vector field through a low-pass filter equivalence analytical model, the cutoff frequency and characteristic frequencies of the lattice radiotherapy dose distribution are obtained, enabling the deviation between the planned dose and the cumulative dose to be attributed and quantitatively explained at the frequency domain level; by obtaining peak-valley dose index acquisition modules, peak diffusion quantification index, valley filling quantification index, and peak-valley dose degradation rate comprehensive index are obtained, achieving precise quantification of the peak-valley structural distortion mode of lattice radiotherapy; by using a frequency domain physical prediction model to fit the characteristic frequencies and cutoff frequencies of the lattice radiotherapy dose distribution, the upper limit of the progressive degradation rate and the degradation curve steepness parameters are determined, obtaining the predicted peak-valley dose degradation rate, realizing prospective quantitative calculation of the dose degradation degree based on the individual patient's motion amplitude; by using a quality judgment module to automatically output quality judgment results based on the predicted peak-valley dose degradation rate, the automation and standardization of dose distortion assessment are achieved.

[0020] Furthermore, in the consistent spatial segmentation radiotherapy dose distortion quantification system based on low-pass filtering modeling, the data preprocessing module includes a data reading module and a data alignment module;

[0021] The data reading module is used to read physical dose data and geometric spatial data; the physical dose data includes a static planned dose distribution matrix and a cumulative dose distribution matrix; the geometric spatial data includes the gross tumor target area contour, the lattice high-dose target area contour, the lattice low-dose valley area contour, and the organ at risk contour.

[0022] The data alignment module is used to unify the coordinate system and grid resolution of the static planned dose distribution matrix and the cumulative dose distribution matrix to obtain preprocessed physical dose data.

[0023] In the above scheme, by reading physical dose data, the static planned dose distribution matrix and the cumulative dose distribution matrix are obtained, providing two sets of raw numerical data sources for dose analysis within the system. By reading geometric spatial data, the contours of the gross tumor target area, the high-dose target area of ​​the lattice, and the organs at risk are obtained. The contour of the low-dose valley area of ​​the lattice is automatically generated by the spatial logical difference between the contours of the gross tumor target area and the high-dose target area of ​​the lattice, without the need for manual drawing. The geometric data defines the calculation area of ​​each quantitative index. The coordinate system and grid resolution of the static planned dose matrix and the cumulative dose matrix are unified by the data alignment module, so that the two sets of dose data correspond precisely on a voxel-by-voxel basis in three-dimensional space, ensuring the uniformity of spatial position.

[0024] Furthermore, a consistent spatial segmentation radiotherapy dose distortion quantification system based on low-pass filtering modeling is provided, wherein the low-pass filtering equivalence analytical model includes a cumulative operator definition module, an equivalence proof module, and a frequency domain modulation transfer function analysis module.

[0025] The cumulative operator definition module uses the preprocessed cumulative dose distribution matrix to establish a mathematical generation model for the cumulative dose distribution matrix;

[0026] The equivalence proof module defines a spatial weight kernel function based on the mathematical generation model of the cumulative dose distribution matrix. The spatial weight kernel function is used to rewrite the mathematical generation model of the cumulative dose distribution matrix into a convolutional form. Due to the mathematical properties of the spatial weight kernel function, the convolutional form of the cumulative dose distribution matrix is ​​equivalent to a three-dimensional spatial low-pass filter.

[0027] The frequency domain modulation transfer function analysis module performs a Fourier transform on the convolutional cumulative dose distribution matrix to establish the modulation transfer function of the low-pass filter. It then approximates the modulation transfer function of the low-pass filter using a Gaussian distribution to obtain the cutoff frequency and the characteristic frequency of the lattice radiotherapy dose distribution.

[0028] In the above scheme, a mathematical generation model is established using the preprocessed cumulative dose distribution matrix through the cumulative operator definition module, transforming the multi-temporal dose accumulation process under respiratory motion into an analytical mathematical expression. The equivalence proof module constructs a spatial weight kernel function to rewrite the mathematical generation model into a convolutional form. Utilizing the three mathematical properties of the weight kernel function—finite support domain, normalization condition, and spatial attenuation—it is proven that the convolutional form of the cumulative dose distribution matrix is ​​equivalent to a three-dimensional spatial low-pass filter. This transforms the dose accumulation calculation process from point-by-point numerical calculation in the spatial domain into a low-pass filtering model, revealing the mathematical essence of the structural influence of the mathematical generation model on the dose distribution. Finally, the frequency domain modulation transfer function analysis module performs a Fourier transform on the convolutional form to establish the modulation transfer function. An analytical expression for the cutoff frequency is obtained by approximating the Gaussian distribution, enabling quantitative description of dose distortion in the frequency domain and providing frequency domain characteristic parameters for the construction of a frequency domain physical prediction model.

[0029] Furthermore, in the consistent spatial segmentation radiotherapy dose distortion quantization system based on low-pass filtering modeling, the mathematical properties of the spatial weight kernel function include finite support domain, normalization condition, and spatial attenuation.

[0030] The finite support domain indicates that the tissue displacement caused by respiratory motion has a finite range in three-dimensional coordinates;

[0031] The normalization condition indicates that the integral of the spatial weight kernel function over all three-dimensional coordinates is equal to 1;

[0032] The spatial attenuation indicates that the probability of an extreme case where the displacement amplitude far exceeds the normal range approaches zero.

[0033] In the above scheme, the finite support domain characteristic of the spatial weight kernel function limits the respiratory motion displacement to only take values ​​within a finite range of three-dimensional coordinates, ensuring the boundedness of the spatial weight kernel function in the physical displacement space, which is consistent with the actual motion law of the thoracic and abdominal organs constrained by anatomical structures. The normalization condition ensures that the integral of the weight kernel function in the whole space is equal to 1, giving it the mathematical properties of a probability density function, ensuring the integrity of the displacement probability distribution and the conservation of total probability. The spatial attenuation ensures that the probability of the displacement amplitude far exceeding the normal range approaches zero, eliminating the interference of extreme displacement on the cumulative dose calculation, so that the spatial weight kernel function only affects the dose distribution within the actual motion displacement range. These three mathematical properties together define the mathematical boundary of the spatial weight kernel function as a displacement probability density function.

[0034] Furthermore, in the consistent spatial segmentation radiotherapy dose distortion quantization system based on low-pass filtering modeling, the frequency domain modulation transfer function analysis module includes a frequency domain transformation module, a Gaussian approximation module, and a cutoff frequency analysis module.

[0035] The frequency domain transformation module performs a three-dimensional Fourier transform on the convolutional cumulative dose distribution matrix to establish the modulation transfer function of the low-pass filter;

[0036] When the breathing motion is approximated as a Gaussian distribution, the Gaussian approximation module approximates the spatial weight kernel function as an anisotropic Gaussian function to obtain the modulation transfer function of the low-pass filter.

[0037] The cutoff frequency analysis module obtains the cutoff frequency based on the head-to-toe motion amplitude and the proportionality constant, and determines the characteristic frequency of the lattice radiotherapy dose distribution based on the lattice spacing.

[0038] In the above scheme, a frequency domain transformation module performs a three-dimensional Fourier transform on the convolution form to convert the dose analysis from the spatial domain to the frequency domain, establishing the modulation transfer function of the low-pass filter, thus providing a quantitative description of the frequency domain response characteristics of the low-pass filter. A Gaussian approximation module approximates the spatial weight kernel function as an anisotropic Gaussian function and performs a Fourier transform when respiratory motion is approximated as a Gaussian distribution, obtaining an analytical expression for the modulation transfer function of the low-pass filter, providing a directly analytical mathematical form for calculating the cutoff frequency. A cutoff frequency analysis module obtains the cutoff frequency based on the amplitude of head-to-toe motion and the proportionality constant, and determines the characteristic frequency of the lattice radiotherapy dose distribution based on the lattice spacing, obtaining two frequency domain characteristic parameters used to determine whether the characteristic frequency of the lattice radiotherapy dose distribution is attenuated by the low-pass filter.

[0039] Furthermore, in the consistent spatial segmentation radiotherapy dose distortion quantification system based on low-pass filtering modeling, the peak-valley dose index acquisition module includes a peak area diffusion quantification index acquisition module, a valley area filling quantification index acquisition module, and a peak-valley dose degradation rate comprehensive index acquisition module. The peak region dispersion quantization index acquisition module uses the lattice high-dose target region contour, the preprocessed static planned dose distribution matrix and the cumulative dose distribution matrix to obtain the static peak dose and the cumulative peak dose, and uses the static peak dose and the cumulative peak dose to obtain the peak region dispersion quantization index. The valley filling quantification index acquisition module uses the lattice low-dose valley contour, the preprocessed static planned dose distribution matrix and the cumulative dose distribution matrix to obtain the static valley dose and the cumulative valley dose, and uses the static valley dose and the cumulative valley dose to obtain the valley filling quantification index.

[0040] The peak-valley dose degradation rate comprehensive index acquisition module obtains the static planned peak-valley dose ratio based on the static peak dose and static valley dose, obtains the cumulative peak-valley dose ratio based on the cumulative peak dose and cumulative valley dose, and obtains the peak-valley dose degradation rate comprehensive index using the static planned peak-valley dose ratio and the cumulative peak-valley dose ratio.

[0041] In the above scheme, the peak dispersion quantification index acquisition module spatially averages the static planned dose and cumulative dose within the high-dose target region contour of the lattice to obtain the attenuation degree of peak dose from the planned value to the cumulative value, thus quantitatively characterizing the dose loss in the high-dose peak region; the valley filling quantification index acquisition module spatially averages the static planned dose and cumulative dose within the low-dose valley region contour of the lattice to obtain the increase degree of valley dose from the planned value to the cumulative value, thus quantitatively characterizing the dose filling in the low-dose valley region; the peak-valley dose degradation rate comprehensive index acquisition module first obtains the static planned peak-valley dose ratio and the cumulative peak-valley dose ratio from the ratio of peak dose to valley dose, and then obtains the peak-valley dose degradation rate comprehensive index from the relative change of the two, thus integrating the combined destruction degree of peak-valley contrast by the two effects of peak dispersion and valley filling into a single quantitative value characterization.

[0042] Furthermore, in the consistent spatial segmentation radiotherapy dose distortion quantification system based on low-pass filtering modeling, the peak region diffusion quantification index acquisition module includes a static peak dose calculation module, a cumulative peak dose calculation module, and a peak region diffusion quantification index calculation module.

[0043] The static peak dose calculation module performs spatial averaging on the preprocessed static planned dose distribution matrix within the lattice high-dose target region to obtain the static peak dose.

[0044] The cumulative peak dose calculation module performs spatial averaging on the preprocessed cumulative dose distribution matrix within the high-dose target region of the lattice to obtain the cumulative peak dose.

[0045] The peak region diffusion quantification index calculation module obtains the peak region diffusion quantification index based on the static peak dose and the cumulative peak dose.

[0046] In the above scheme, the static peak dose calculation module spatially averages the static planned dose distribution matrix within the high-dose target area of ​​the lattice to obtain a single value of the static peak dose, so that the overall planned dose level of the high-dose target area of ​​the lattice is output in scalar form; the cumulative peak dose calculation module spatially averages the cumulative dose distribution matrix within the high-dose target area of ​​the lattice to obtain a single value of the cumulative peak dose, so that the overall actual dose level of the high-dose target area of ​​the lattice after the influence of respiratory motion is output in scalar form; the peak area diffusion quantification index calculation module calculates the peak area diffusion quantification index based on the relative change of the static peak dose and the cumulative peak dose, so that the degree of attenuation of the peak area dose from the planned value to the cumulative value is represented by a single value.

[0047] Furthermore, in the consistent spatial segmentation radiotherapy dose distortion quantification system based on low-pass filtering modeling, the valley filling quantification index acquisition module includes a static valley dose calculation module, a cumulative valley dose calculation module, and a valley filling quantification index calculation module.

[0048] The static valley dose calculation module performs spatial averaging on the preprocessed static planned dose distribution matrix within the low-dose valley region of the lattice to obtain the static valley dose.

[0049] The cumulative valley dose calculation module performs spatial averaging on the preprocessed cumulative dose distribution matrix within the low-dose valley region of the lattice to obtain the cumulative valley dose.

[0050] The valley filling quantification index calculation module obtains the valley filling quantification index based on the static valley dose and the cumulative valley dose.

[0051] In the above scheme, the static valley dose calculation module spatially averages the static planned dose distribution matrix within the low-dose valley region of the lattice to obtain a single value of the static valley dose, so that the overall planned dose level of the low-dose valley region of the lattice is output in scalar form; the cumulative valley dose calculation module spatially averages the cumulative dose distribution matrix within the low-dose valley region of the lattice to obtain a single value of the cumulative valley dose, so that the overall actual dose level of the low-dose valley region of the lattice after respiratory motion is output in scalar form; the valley filling quantification index calculation module calculates the valley filling quantification index based on the relative change of the static valley dose and the cumulative valley dose, so that the degree of filling of the valley dose from the planned value to the cumulative value is represented by a single value.

[0052] Furthermore, in the consistent spatial segmentation radiotherapy dose distortion quantization system based on low-pass filtering modeling, the frequency domain physical prediction model is defined by the following formula:

[0053] ;

[0054] in, For frequency domain physical prediction models, For the predicted peak-to-trough dose degradation rate, This is the upper limit of the progressive degradation rate. It is a natural exponential function. This is the parameter for the steepness of the degradation curve. The cutoff frequency, The characteristic frequency of the lattice radiotherapy dose distribution.

[0055] In the above scheme, a nonlinear mapping relationship between the ratio of the characteristic frequency and cutoff frequency of the lattice radiotherapy dose distribution and the peak-valley dose degradation rate is established through a frequency domain physical prediction model. After inputting the characteristic frequency and cutoff frequency of the lattice radiotherapy dose distribution, the predicted peak-valley dose degradation rate is directly output, so that the dose degradation degree can be quantitatively predicted during the planning stage without waiting for post-dose calculation. The nonlinear function form is used to describe the law of the peak-valley dose degradation rate changing with the frequency ratio. The upper limit of the asymptotic degradation rate limits the convergence boundary of the degradation rate when the motion amplitude is extremely large. The degradation curve steepness parameter controls the rate of increase of the degradation rate with the frequency ratio. Compared with the traditional linear model, it can accurately describe the physical law that the peak-valley dose degradation rate tends to saturate when the motion amplitude increases. It still maintains prediction accuracy under large motion amplitude conditions and solves the problem of large prediction deviation of the traditional linear model in the large amplitude range.

[0056] Furthermore, in the consistent spatial segmentation radiotherapy dose distortion quantification system based on low-pass filtering modeling, the quality judgment result is specifically as follows:

[0057] when At the same time, the peak-to-valley dose ratio of the lattice radiotherapy plan remained good, and the plan quality was excellent;

[0058] when At that time, the peak-to-trough dose ratio of the lattice radiotherapy plan showed a slight degradation, and the plan quality was acceptable;

[0059] when At that time, the peak-to-valley dose ratio of the lattice radiotherapy plan showed moderate degradation, and the quality of the plan needed to be monitored;

[0060] when At that time, the peak-to-valley dose ratio of the lattice radiotherapy plan deteriorated significantly, and the plan quality failed to meet the standards.

[0061] In the above scheme, the quality judgment module compares the predicted peak-valley dose degradation rate with the preset four-level threshold range and directly outputs the corresponding quality judgment result without the need for manual interpretation of each case. Based on the predicted peak-valley dose degradation rate range, it corresponds to four levels: excellent plan quality, acceptable plan quality, plan quality requiring attention, and plan quality failing to meet standards. This presents the dose distortion assessment results in a standardized graded conclusion, providing a unified judgment basis for different quality levels and avoiding the subjective differences and inconsistent standards caused by manual interpretation. Attached Figure Description

[0062] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0063] Figure 1 This is a structural diagram of a spatial segmentation radiotherapy dose distortion quantization system based on low-pass filtering modeling.

[0064] Figure 2 This is a static planned dose distribution map.

[0065] Figure 3 This is a cumulative dose distribution map.

[0066] Figure 4 This is an analytical diagram of the frequency domain modulation transfer function.

[0067] Figure 5 This is a comparison chart of the frequency domain physics prediction model and the auxiliary linear regression model.

[0068] Figure 6 This is a frequency domain analysis diagram. Detailed Implementation

[0069] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0070] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this invention, the terms "first," "second," etc., are used only for distinguishing descriptions and should not be construed as indicating or implying relative importance, or suggesting any such actual relationship or order between these entities or operations. Additionally, the terms "connected," "linked," etc., can refer to a direct connection between elements or an indirect connection via other elements.

[0071] This invention is achieved through the following technical solutions, such as... Figure 1 As shown, a spatial segmentation radiotherapy dose distortion quantification system based on low-pass filtering modeling includes a data preprocessing module, a low-pass filtering equivalence analytical model, a peak-valley dose index acquisition module, a frequency domain physical prediction model, and a quality judgment module.

[0072] The data preprocessing module is used to read physical dose data and geometric space data, preprocess the physical dose data, and obtain preprocessed physical dose data.

[0073] The data preprocessing module includes a data reading module and a data alignment module.

[0074] The data reading module is used to read physical dose data and geometric spatial data; the physical dose data includes a static planned dose distribution matrix and a cumulative dose distribution matrix; the geometric spatial data includes gross tumor target volume (GTV), lattice high-dose target volume (GTV-Lat), lattice low-dose valley (GTV-Valley), and organ at risk (OARs).

[0075] It should be noted that the low-dose valley region contour of the lattice is automatically generated based on the spatial logical difference between the gross tumor target region contour and the high-dose target region contour of the lattice.

[0076] like Figures 2-3 As shown, the data alignment module is used to unify the coordinate system and grid resolution (e.g., 2×2×2 mm³) of the static planned dose distribution matrix and the cumulative dose distribution matrix to obtain preprocessed physical dose data.

[0077] In this embodiment, data from 31 patients with thoracic and abdominal tumors were collected, specifically 12 cases of liver cancer, 11 cases of lung cancer, and 8 cases of pancreatic cancer. The range of head and foot movements in each patient with thoracic and abdominal tumors was recorded. The range was 5-45 mm, with a median of 18 mm. Data from 31 patients with thoracic and abdominal tumors were divided into a model fitting set of 21 cases and an independent validation set of 10 cases.

[0078] The low-pass filter equivalence analytical model is used to perform mathematical equivalence proof and frequency domain analysis using preprocessed physical dose data to obtain the cutoff frequency and the characteristic frequency of the lattice radiotherapy dose distribution.

[0079] The low-pass filter equivalence analytical model includes a cumulative operator definition module, an equivalence proof module, and a frequency domain modulation transfer function analysis module.

[0080] The cumulative operator definition module uses the preprocessed cumulative dose distribution matrix to establish a mathematical generation model for the cumulative dose distribution matrix, the formula of which is:

[0081] ;

[0082] ;

[0083] in, For the preprocessed three-dimensional coordinates The cumulative dose distribution matrix, It is a three-dimensional coordinate vector. The location bin index for discretizing the respiratory motion trajectory. The total number of position bins for discretizing the respiratory motion trajectory. Let be the dwell time probability of the box at position k. For the k-th location box, the three-dimensional coordinates Instantaneous dose, For the k-th location box, the three-dimensional coordinates A three-dimensional vector field, For left and right indexes in space, For spatial forward and backward indexing, This provides a spatial head-to-foot orientation index.

[0084] In this embodiment, the respiratory trajectory is discretized into a position box (from end of expiration to end of inspiration) with K=5. When k=1, the probability of dwell time at end of expiration is the highest. This reflects the physiological characteristic of having the longest end-expiratory residence time in normal breathing; when k=2 or 3, the probability of residence time in the transition phase and the intermediate phase of expiration is equal. This reflects the physiological characteristics of the transition phase from expiration to inspiration, with faster velocity and shorter residence time in the intermediate phase. When k=4, the residence time probability of the inspiratory transition phase is the lowest. This reflects the physiological characteristics of the respiratory movement being fastest and the residence time shortest during the acceleration process from the middle phase to the end of inspiration. When k=5, the residence time at the end of inspiration is lower than that at the end of expiration, but higher than that at other phases. This reflects the physiological characteristic that after the lung tissue has fully expanded at the end of inspiration, the diaphragm is in a state of tension but still has a brief pause. It satisfies the normalization condition.

[0085] The equivalence verification module defines a spatial weight kernel function based on the mathematical generation model of the cumulative dose distribution matrix. Using this spatial weight kernel function, the mathematical generation model of the cumulative dose distribution matrix is ​​rewritten in convolutional form, as shown in the formula:

[0086] ;

[0087] ;

[0088] in, For three-dimensional displacement Spatial weight kernel function, It is a three-dimensional displacement vector. For Dirac delta function, For the preprocessed three-dimensional coordinates The static planned dose distribution matrix, For three-dimensional convolution operations, Three-dimensional coordinates Spatial weight kernel function.

[0089] It is important to note that the spatial weight kernel function describes the probability distribution of displacement caused by respiratory motion in a physical sense.

[0090] The weighting kernel function has three mathematical properties; these mathematical properties are as follows:

[0091] (1) Finite support domain: , ;

[0092] in, For the support set of the spatial weight kernel function, For the maximum displacement amplitude, To maximize the value of index k for all location bins, This is a modulo operation.

[0093] It is important to note that a finite support domain indicates that the tissue displacement caused by respiratory motion has a finite range in three-dimensional coordinates.

[0094] (2) Normalization conditions: ;

[0095] It is important to note that the normalization condition means that the integral of the spatial weight kernel function over all three-dimensional coordinates is equal to 1, which guarantees the probability density function property of the weight kernel function.

[0096] (3) Spatial attenuation: when , ;

[0097] It is important to note that spatial attenuation means that as the magnitude of the three-dimensional displacement vector approaches infinity, the value of the spatial weight kernel function approaches 0, meaning that the probability of extreme cases where the displacement amplitude far exceeds the normal range approaches zero.

[0098] Due to the mathematical properties of the spatial weight kernel function, the cumulative dose distribution matrix in convolutional form is equivalent to a three-dimensional spatial low-pass filter, as shown in the formula:

[0099] ;

[0100] in, For constant equivalent, This is a three-dimensional low-pass filter operator.

[0101] The frequency domain modulation transfer function analysis module performs a Fourier transform on the convolutional cumulative dose distribution matrix to establish the modulation transfer function of the low-pass filter. It then approximates the modulation transfer function of the low-pass filter using a Gaussian distribution to obtain the cutoff frequency.

[0102] like Figure 4 As shown, the frequency domain modulation transfer function analysis module includes a frequency domain transformation module, a Gaussian approximation module, and a cutoff frequency analysis module.

[0103] The frequency domain transformation module performs a three-dimensional Fourier transform on the convolutional cumulative dose distribution matrix to establish the modulation transfer function of the low-pass filter, as shown in the following formula:

[0104] ;

[0105] ;

[0106] ;

[0107] Three-dimensional spatial frequency The three-dimensional Fourier spectrum of the cumulative dose distribution matrix at that location. It is a three-dimensional frequency vector. Three-dimensional spatial frequency The three-dimensional Fourier spectrum of the static planned dose distribution matrix. The modulation transfer function (three-dimensional frequency) of the low-pass filter. (Three-dimensional Fourier spectrum of the spatial weight kernel function). For three-dimensional Fourier transform operators, The spatial frequency is in the left-right direction. For the spatial frequency in the forward and backward directions, The spatial frequency is in the head-to-feet direction.

[0108] It is important to note that, due to the mathematical properties of the weight kernel function, In the low frequency band Maintain high response in the high frequency band Significant attenuation, which is a low-pass response function.

[0109] in, This is the cutoff frequency.

[0110] It is important to note that the cutoff frequency is specifically the spatial frequency at which the modulation transfer function decays to -3dB (approximately 0.707). ,but .

[0111] When the breathing motion is approximated as a Gaussian distribution, the Gaussian approximation module approximates the spatial weight kernel function as an anisotropic Gaussian function, thus obtaining the modulation transfer function of the low-pass filter, as shown in the formula:

[0112] ;

[0113] ;

[0114] in, The standard deviation of the left and right respiratory movement displacement distribution. The standard deviation of the displacement distribution of the preceding and following respiratory movements. This represents the standard deviation of the head-to-foot respiratory displacement distribution.

[0115] In this embodiment, the equivalence of the low-pass filter was verified: the cumulative dose distribution matrix obtained by using a mathematical generation model of the cumulative dose distribution matrix and a convolutional form of the cumulative dose distribution matrix from data of 31 patients with thoracic and abdominal tumors was compared, and the maximum relative error was < 0.1%. The modulation transfer function obtained by analytical derivation of the frequency domain transformation module was compared with the modulation transfer function obtained by directly performing a fast Fourier transform on the spatial weight kernel function, and the correlation coefficient was compared. Therefore, the equivalence of low-pass filtering has been rigorously verified at the numerical level.

[0116] It is important to note that The magnitude of the motion is proportional to the direction of each component, for example, the uniform distribution approximation: , , For the directional motion amplitude, z = AP, LR, SI, for example, a sinusoidal distribution approximation: .

[0117] Specifically, the directional motion amplitude is extracted from a three-dimensional vector field, traversing all position boxes k and all spatial position three-dimensional coordinates. The maximum absolute value of the displacement of the three-dimensional vector field in three directions is extracted to obtain the motion amplitude in the corresponding direction.

[0118] The cutoff frequency analysis module obtains the cutoff frequency based on the head-to-toe motion amplitude and the proportionality constant, and determines the characteristic frequency of the lattice radiotherapy dose distribution based on the lattice spacing, using the following formula:

[0119] ;

[0120] ;

[0121] in, It is a proportionality constant. This refers to the range of motion in the head-to-toe direction (for tumors in the chest and abdomen, the range of motion in the left-to-right direction is also considered). and amplitude of forward and backward movement (Usually small, negligible) These are the characteristic frequencies of the lattice radiotherapy dose distribution. is the lattice spacing.

[0122] Specifically, when This indicates a severe degradation in the peak-to-trough dose ratio;

[0123] when At that time, it indicates that the peak-to-trough dose ratio has not deteriorated significantly.

[0124] In this embodiment, a retrospective case of liver cancer was selected. mm, mm, ,but , ,but , This indicates a severe degradation in the peak-to-trough dose ratio.

[0125] The peak-valley dose index acquisition module obtains a comprehensive index of peak-valley dose degradation rate based on geometric spatial data and preprocessed physical dose data.

[0126] The peak-valley dose index acquisition module includes a peak region diffusion quantification index acquisition module, a valley region filling quantification index acquisition module, and a peak-valley dose degradation rate comprehensive index acquisition module.

[0127] The peak region dispersion quantization index acquisition module uses the lattice high-dose target region contour, the preprocessed static planned dose distribution matrix and the cumulative dose distribution matrix to obtain the static peak dose and the cumulative peak dose, and uses the static peak dose and the cumulative peak dose to obtain the peak region dispersion quantization index.

[0128] The peak region diffusion quantification index acquisition module includes a static peak dose calculation module, a cumulative peak dose calculation module, and a peak region diffusion quantification index calculation module.

[0129] The static peak dose calculation module performs spatial averaging on the preprocessed static planned dose distribution matrix within the lattice high-dose target region to obtain the static peak dose, as shown in the formula:

[0130] ;

[0131] in, This is the static peak dose. The volume of the high-dose target region in the crystal lattice. The outline of the high-dose target region in the crystal lattice.

[0132] The cumulative peak dose calculation module performs spatial averaging on the preprocessed cumulative dose distribution matrix within the high-dose target region of the lattice to obtain the cumulative peak dose, as shown in the formula:

[0133] ;

[0134] in, This represents the cumulative peak dose.

[0135] The peak region diffusion quantification index calculation module obtains the peak region diffusion quantification index based on the static peak dose and cumulative peak dose, using the following formula:

[0136] ;

[0137] in, This is a quantitative indicator for peak region dispersion.

[0138] The valley filling quantification index acquisition module obtains static valley dose and cumulative valley dose by using the lattice low-dose valley contour, the preprocessed static planned dose distribution matrix and cumulative dose distribution matrix, and obtains valley filling quantification index using static valley dose and cumulative valley dose.

[0139] The valley filling quantitative index acquisition module includes a static valley dose calculation module, a cumulative valley dose calculation module, and a valley filling quantitative index calculation module.

[0140] The static valley dose calculation module performs spatial averaging on the preprocessed static planned dose distribution matrix within the low-dose valley region of the lattice to obtain the static valley dose, as shown in the formula:

[0141] ;

[0142] in, This is the static trough dose. The volume of the low-dose valley region of the crystal lattice. The outline of the low-dose valley region of the crystal lattice.

[0143] The cumulative valley dose calculation module performs spatial averaging on the preprocessed cumulative dose distribution matrix within the low-dose valley region of the lattice to obtain the cumulative valley dose, as shown in the formula:

[0144] ;

[0145] in, This is the cumulative trough dose.

[0146] The valley filling quantification index calculation module obtains the valley filling quantification index based on static valley dose and cumulative valley dose, using the following formula:

[0147] ;

[0148] in, Quantitative indicators for filling in valley areas.

[0149] The peak-valley dose degradation rate comprehensive index acquisition module obtains the static planned peak-valley dose ratio based on the static peak dose and static valley dose, and the cumulative peak-valley dose ratio based on the cumulative peak dose and cumulative valley dose. Using the static planned peak-valley dose ratio and the cumulative peak-valley dose ratio, the peak-valley dose degradation rate comprehensive index is obtained, with the following formula:

[0150] ;

[0151] ;

[0152] ;

[0153] in, This is the static peak-to-trough dose ratio. For the cumulative peak-to-trough dose ratio, It is a comprehensive index of peak and valley dose degradation rate.

[0154] In this embodiment, a retrospective case of liver cancer was selected. mm, , ,but , , ,but ,final This indicates that respiratory movements no longer significantly differentiate the dose between the peak and trough regions, severely disrupting the peak-trough dose ratio.

[0155] In this embodiment, the peak diffusion quantification index, valley filling quantification index, and peak-valley dose degradation rate comprehensive index were calculated for data from 31 patients with thoracic and abdominal tumors. The numerical range and median of each index were statistically analyzed, and correlation analysis was performed with the amplitude of head-to-toe movement. The results are shown in Table 1.

[0156] Table 1: Performance of the three major quantitative indicators:

[0157] ;

[0158] Where P is the significance test probability value.

[0159] Therefore, it can be seen that all three quantitative indicators are strongly positively correlated with the amplitude of respiratory motion, which can effectively quantify the degree of dose distortion of lattice radiotherapy caused by respiratory motion. The comprehensive indicator of peak-valley dose degradation rate has the highest correlation coefficient, and it is reasonable to use it as the core indicator for quality control of lattice radiotherapy.

[0160] The frequency domain physical prediction model is used to fit and determine the upper limit of the progressive degradation rate and the steepness parameter of the degradation curve based on the characteristic frequency and cutoff frequency of the lattice radiotherapy dose distribution, and to obtain the predicted peak-valley dose degradation rate, as shown in the formula:

[0161] ;

[0162] in, This is a frequency domain physical prediction model, where represents the predicted peak-to-valley dose degradation rate. This is the upper limit of the progressive degradation rate. It is a natural exponential function. This is the parameter for the steepness of the degradation curve.

[0163] In the embodiments, Typical values ​​are 50-60%. Typical values ​​are 1.5-2.5.

[0164] In this embodiment, the amplitude of head-to-foot movement was extracted from data of 31 patients with thoracic and abdominal tumors. Combined index of peak and trough dose degradation rate The statistical distribution of head-to-foot movement amplitude is as follows: minimum value is 5 mm, 25th percentile is 12 mm, median is 18 mm, 75th percentile is 28 mm, and maximum value is 45 mm, covering a wide range of movement amplitude from mild to severe.

[0165] in, The range of head-to-toe movement for the j-th patient is given. is a composite index of peak-to-trough dose degradation rate for patient j, where j is the index of patients with thoracic and abdominal tumors.

[0166] The Levenberg-Marquardt nonlinear least squares method was used to fit a dataset of 21 models to establish a frequency domain physical prediction model. The values ​​of the upper limit of the gradual degradation rate and the steepness parameter of the degradation curve were obtained, as shown in the following formulas:

[0167] ;

[0168] It can be simplified to:

[0169] ;

[0170] The frequency domain physical prediction model was triple-validated using 10 independent validation sets: the triple validation included the coefficient of determination, the Shapiro-Wilk test, and the Breusch-Pagan test.

[0171] Specifically, the coefficient of determination The p-value of the Shapiro-Wilk test The p-value of the Breusch-Pagan test If all conditions are met, the verification is passed.

[0172] like Figure 5 As shown, the frequency domain physical prediction model is evaluated as This indicates that the frequency domain physical prediction model can explain 94% of the data. Mutations, It satisfies the normality assumption. It satisfies the homoscedasticity assumption, with a mean absolute error (MAE) of 1.8%, meeting the accuracy requirements for clinical applications.

[0173] As a benchmark, an auxiliary linear regression model is also established, with the following formula:

[0174] ;

[0175] in, The slope of the linear regression. This is the intercept for linear regression.

[0176] The specific data for patients with thoracic and abdominal tumors are as follows: .

[0177] Coefficient of determination in auxiliary linear regression model The mean absolute error is 2.3%, and the relative error is 6.8%. The coefficient of determination of the frequency domain physical prediction model. The mean absolute error was 1.8%, and the relative error was 5.2%, all three indicators were better than the auxiliary linear regression model.

[0178] It is evident that the frequency domain physical prediction model is significantly superior to the traditional linear model. Especially in the large motion amplitude range, the frequency domain physical prediction model exhibits better nonlinear fitting ability and more accurately describes the physical law that the peak-valley dose degradation rate tends to saturate when the motion amplitude increases.

[0179] The quality assessment module is used to obtain quality assessment results based on the predicted peak-valley dose degradation rate:

[0180] when At the same time, the peak-to-valley dose ratio of the lattice radiotherapy plan remained good, and the plan quality was excellent;

[0181] when At that time, the peak-to-trough dose ratio of the lattice radiotherapy plan showed a slight degradation, and the plan quality was acceptable;

[0182] when At that time, the peak-to-valley dose ratio of the lattice radiotherapy plan showed moderate degradation, and the quality of the plan needed to be monitored;

[0183] when At that time, the peak-to-valley dose ratio of the lattice radiotherapy plan deteriorated significantly, and the plan quality failed to meet the standards.

[0184] In some embodiments, the quality control module can also generate differential dose maps and such as Figure 6 The frequency domain analysis diagram shown is used as an auxiliary analysis.

[0185] In this embodiment, the consistency verification of quality control judgment is performed as follows: To verify the consistency between the automatic judgment result of the quality control module of the present invention and the manual judgment, 10 independent validation set patient data were selected. The quality control judgment of the present invention and three senior medical physicists (physicist A, physicist B, and physicist C) were used to independently perform quality control judgment. The Kappa consistency coefficient was used as the evaluation index to analyze the consistency between the quality judgment result of the present invention and the quality judgment result of each physicist. The results are shown in the table.

[0186] Table 2: Consistency Verification of Quality Control Judgments

[0187] ;

[0188] The Kappa consistency coefficients between the judgment results of the quality control module of this invention and the judgment results of the three physicists are 0.88, 0.85 and 0.91, respectively, with an average consistency of 0.88. The Kappa consistency coefficients are all greater than 0.80, which confirms that the quality control module of this invention can be used as an objective and automated quality control tool to replace the manual judgment process.

[0189] In the embodiments, the ablation experiment: the comprehensive index of peak and valley dose degradation rate was used as the prediction target, and the coefficient of determination and mean absolute error were used as evaluation indicators to compare the prediction performance of four model variants, as shown in Table 3.

[0190] Table 3: Ablation Experiment:

[0191] ;

[0192] By adding a comprehensive peak-valley dose degradation rate index and a frequency domain physical prediction model to the dual-quantitative index, the coefficient of determination was further improved to 0.94, and the mean absolute error was reduced to 1.8%. The results show that the addition of the frequency domain physical prediction model improves the prediction accuracy by about 25% compared to the single spatial index.

[0193] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A spatial segmentation radiotherapy dose distortion quantization system based on low-pass filter modeling, characterized in that, It includes a data preprocessing module, a low-pass filter equivalence analytical model, a peak-valley dose index acquisition module, a frequency domain physical prediction model, and a quality assessment module; The data preprocessing module is used to read physical dose data and geometric space data, preprocess the physical dose data, and obtain preprocessed physical dose data. The low-pass filter equivalence analytical model is used to perform mathematical equivalence proof and frequency domain analysis using preprocessed physical dose data to obtain the characteristic frequency of the cutoff frequency lattice radiotherapy dose distribution. The peak-valley dose index acquisition module obtains a comprehensive index of peak-valley dose degradation rate based on geometric spatial data and preprocessed physical dose data. The frequency domain physical prediction model is used to fit and determine the upper limit of the progressive degradation rate and the steepness parameter of the degradation curve based on the characteristic frequency and cutoff frequency of the lattice radiotherapy dose distribution, and to obtain the predicted peak and valley dose degradation rate. The quality assessment module is used to obtain quality assessment results based on the predicted peak and valley dose degradation rates.

2. The spatial segmentation radiotherapy dose distortion quantization system based on low-pass filter modeling according to claim 1, characterized in that, The data preprocessing module includes a data reading module and a data alignment module; The data reading module is used to read physical dose data and geometric spatial data; the physical dose data includes a static planned dose distribution matrix and a cumulative dose distribution matrix; the geometric spatial data includes the gross tumor target area contour, the lattice high-dose target area contour, the lattice low-dose valley area contour, and the organ at risk contour; The data alignment module is used to unify the coordinate system and grid resolution of the static planned dose distribution matrix and the cumulative dose distribution matrix to obtain preprocessed physical dose data.

3. The spatial segmentation radiotherapy dose distortion quantization system based on low-pass filter modeling according to claim 1, characterized in that, The low-pass filter equivalence analytical model includes a cumulative operator definition module, an equivalence proof module, and a frequency domain modulation transfer function analysis module. The cumulative operator definition module uses the preprocessed cumulative dose distribution matrix to establish a mathematical generation model for the cumulative dose distribution matrix; The equivalence proof module defines a spatial weight kernel function based on the mathematical generation model of the cumulative dose distribution matrix. The spatial weight kernel function is used to rewrite the mathematical generation model of the cumulative dose distribution matrix into a convolutional form. Due to the mathematical properties of the spatial weight kernel function, the convolutional form of the cumulative dose distribution matrix is ​​equivalent to a three-dimensional spatial low-pass filter. The frequency domain modulation transfer function analysis module performs a Fourier transform on the convolutional cumulative dose distribution matrix to establish the modulation transfer function of the low-pass filter. It then approximates the modulation transfer function of the low-pass filter using a Gaussian distribution to obtain the cutoff frequency and the characteristic frequency of the lattice radiotherapy dose distribution.

4. The spatial segmentation radiotherapy dose distortion quantization system based on low-pass filter modeling according to claim 3, characterized in that, The mathematical properties of the spatial weight kernel function include finite support domain, normalization condition, and spatial decay. The finite support domain indicates that the tissue displacement caused by respiratory motion has a finite range in three-dimensional coordinates; The normalization condition indicates that the integral of the spatial weight kernel function over all three-dimensional coordinates is equal to 1; The spatial attenuation indicates that the probability of an extreme case where the displacement amplitude far exceeds the normal range approaches zero.

5. The spatial segmentation radiotherapy dose distortion quantization system based on low-pass filter modeling according to claim 3, characterized in that, The frequency domain modulation transfer function analysis module includes a frequency domain transformation module, a Gaussian approximation module, and a cutoff frequency analysis module. The frequency domain transformation module performs a three-dimensional Fourier transform on the convolutional cumulative dose distribution matrix to establish the modulation transfer function of the low-pass filter. When the breathing motion is approximated as a Gaussian distribution, the Gaussian approximation module approximates the spatial weight kernel function as an anisotropic Gaussian function to obtain the modulation transfer function of the low-pass filter. The cutoff frequency analysis module obtains the cutoff frequency based on the head-to-toe motion amplitude and the proportionality constant, and determines the characteristic frequency of the lattice radiotherapy dose distribution based on the lattice spacing.

6. The spatial segmentation radiotherapy dose distortion quantization system based on low-pass filter modeling according to claim 2, characterized in that, The peak-valley dose index acquisition module includes a peak region diffusion quantification index acquisition module, a valley region filling quantification index acquisition module, and a peak-valley dose degradation rate comprehensive index acquisition module. The peak region dispersion quantization index acquisition module obtains the peak region dispersion quantization index by using the lattice high-dose target region contour, the preprocessed static planned dose distribution matrix and the cumulative dose distribution matrix, and obtains the static peak dose and the cumulative peak dose. The valley filling quantification index acquisition module obtains valley filling quantification index by using the lattice low-dose valley contour, the preprocessed static planned dose distribution matrix and the cumulative dose distribution matrix, and obtains static valley dose and cumulative valley dose. The peak-valley dose degradation rate comprehensive index acquisition module obtains the static planned peak-valley dose ratio based on the static peak dose and static valley dose, obtains the cumulative peak-valley dose ratio based on the cumulative peak dose and cumulative valley dose, and obtains the peak-valley dose degradation rate comprehensive index using the static planned peak-valley dose ratio and the cumulative peak-valley dose ratio.

7. The spatial segmentation radiotherapy dose distortion quantization system based on low-pass filter modeling according to claim 6, characterized in that, The peak region diffusion quantification index acquisition module includes a static peak dose calculation module, a cumulative peak dose calculation module, and a peak region diffusion quantification index calculation module. The static peak dose calculation module performs spatial averaging on the preprocessed static planned dose distribution matrix within the lattice high-dose target region to obtain the static peak dose. The cumulative peak dose calculation module performs spatial averaging on the preprocessed cumulative dose distribution matrix within the high-dose target region of the lattice to obtain the cumulative peak dose. The peak region diffusion quantification index calculation module obtains the peak region diffusion quantification index based on the static peak dose and the cumulative peak dose.

8. The spatial segmentation radiotherapy dose distortion quantization system based on low-pass filter modeling according to claim 6, characterized in that, The valley filling quantitative index acquisition module includes a static valley dose calculation module, a cumulative valley dose calculation module, and a valley filling quantitative index calculation module. The static valley dose calculation module performs spatial averaging on the preprocessed static planned dose distribution matrix within the low-dose valley region of the lattice to obtain the static valley dose. The cumulative valley dose calculation module performs spatial averaging on the preprocessed cumulative dose distribution matrix within the low-dose valley region of the lattice to obtain the cumulative valley dose. The valley filling quantification index calculation module obtains the valley filling quantification index based on the static valley dose and the cumulative valley dose.

9. The spatial segmentation radiotherapy dose distortion quantization system based on low-pass filter modeling according to claim 1, characterized in that, The frequency domain physical prediction model is defined by the following formula: ； in, For frequency domain physical prediction models, For the predicted peak-to-trough dose degradation rate, This is the upper limit of the progressive degradation rate. It is a natural exponential function. This is the parameter for the steepness of the degradation curve. The cutoff frequency, The characteristic frequency of the lattice radiotherapy dose distribution.

10. The spatial segmentation radiotherapy dose distortion quantization system based on low-pass filter modeling according to claim 9, characterized in that, The specific quality assessment result is as follows: when At the same time, the peak-to-valley dose ratio of the lattice radiotherapy plan remained good, and the plan quality was excellent; when At that time, the peak-to-trough dose ratio of the lattice radiotherapy plan showed a slight degradation, and the plan quality was acceptable; when The peak-to-valley dose ratio of the time-lattice radiotherapy plan has moderately deteriorated, and the plan quality needs to be monitored. when At that time, the peak-to-valley dose ratio of the lattice radiotherapy plan deteriorated significantly, and the plan quality failed to meet the standards.

Images